The interest graph is coming. Eight ways to get ready.
Social networks like Facebook start with your friends and let you see what you have in common. Interest graph-based models – Springpad, Pinterest, Get Glue – start with your interests and then let you make connections. It’s less about who you know and more about what you care about.
If you happen to have your Google alerts set up to grab the latest blog posts and articles about Pinterest, you’re stream is pretty well populated these days. Add “Facebook Actions” or “Springpad” or “Svpply” or “Hunch” and it gets even more crowded. Maybe that’s why I don’t dare add queries for Google’s new privacy changes or developments like YouTube’s original channels. It would be more than anyone could possibly bear.
With each passing week, the social web evolves. Now that we’ve supposedly mastered Facebook and Twitter, we’re confronted with Google + and all the new interest graph platforms mentioned above. Are we ready? Do we know what to do? Do we have a strategy in place?
Recent research that Mullen just conducted suggests not. We surveyed 160 CMOs and marketing chiefs to find out where they stood when it came to using social media, monitoring the stream, developing conversation strategy and having a plan for tapping the interest graph.
We were surprised at some of the results.
Marketers remain challenged by social media
While 87 percent of respondents claimed that social media was somewhat or very important to their marketing efforts, most of their efforts remained limited to, or at least focused on Facebook. Nearly 80 percent were committed to the world’s largest social network. But fewer than 20 percent were using Google + and a full 80 percent had no focus at all on a platform like Foursquare.
While ongoing engagement emerged as one primary objective (64.5 percent noted it) marketers declared their number one reason for using social media was to generate awareness (76.8 percent), an objective that beat out both customer support (29.7 percent) and building loyalty (53.5 percent).
Among the more disappointing, but perhaps expected findings was the fact that marketers measure success primarily by how many followers and/or likes they generate (71.6 percent). By comparison, downloads (24.5 percent), share of conversation (25.2) and referrals (35.5) remained far less important. The latter is particularly surprising given the social web’s built in ability to inspire word-of-mouth marketing and the sharing of recommendations.
When it comes to content, marketers continue to think like traditional advertisers. They primarily use social platforms to promote products and offers (67.5 percent) and to deliver updates (64.9 percent). Providing utility (33.1) and offering entertainment (22.7) remain far less important concerns.
Despite the flurry of press coverage on the emerging importance of the interest graph, nearly half or respondents (48.7 percent) never heard of the term “interest graph,” and when they had it explained – the ability to connect with consumers in a more meaningful way by tapping into their interests – only 26.6 percent thought it could be “very useful.”
As for all that buzz around Pinterest, a platform generating page views, user growth and inbound links for the early adopter brands? Close to half of our respondents (42.2 percent) never even heard of it, while barely a sliver (4.5 percent) had started using it.
Perhaps that’s no surprise given that 68.8 percent of marketers surveyed capture no interest graph data at all — not preferences, interests, or intentions.
Finally, while brand stewards aren’t quite overwhelmed with the proliferation of platforms, they (44.2 percent) struggle with one fundamental challenge – where to put their resources.
From the social graph to the interest graph
The last finding surprises no one. Getting social media efforts to deliver hard results and ROI is a challenge for the simple reason that most consumers aren’t there to connect with brands and their advertising messages.
But the interest graph platforms can change that. If marketers can suddenly identify people who’ve raised their hands and virtually asked for a “proposal,” they can more easily connect with people who’ll welcome them.
Every social network knows this is the future. Facebook Actions now allows users to tap into and identify friends’ interests — music, tastes in foods and preferences for movies, books and more. Presumably, if you actually know what friends have good taste in music it will now be easier to call on their recommendations. Actions aren’t perfect, however.
You still have to scroll through the stream and most content isn’t really persistent, meaning if you miss it in the stream it’s gone. It still poses challenges for marketers, too. Check out your own page and refresh it a few times. I guarantee that you’ll find the majority of ads that get served to you are completely irrelevant. But the promise is significant. Facebook will inevitably get better at capturing even more data and presumably allow advertisers to more accurately focus messages.
Foursquare, which our research told us is barely on the radar for most marketers will start making recommendations to its users on where to eat and where to vacation based on past behavior and that of friends. Certainly any hospitality marketer – restaurants, chains, museums and hotels – should at least be exploring the possibilities, if not encouraging user participation.
But all of this is still new. The social graph as we know it is only a few years old while the interest graph has been a topic of discussion for a matter of months. So what does it all mean? For brands, it’s definitely not too late to be early. Marketers can still get in on the ground level. But they need to embrace it and work to leverage it.
For social media practitioners, there’s work to be done. We need to learn, educate each other, experiment and develop effective strategies and tactics.
Eight steps you can take to get ready
- Learn the difference between the social graph and the interest graph. This simple description, by David Rogers writing in Read Write Web might help.*
- Read Grouped and get a better sense of how influence happens on the social web. The Tipping Point is a fallacy. Influence isn’t what you think it is. Small groups are what really matter.
- Open accounts on at least a few of the platforms. We would recommend Pinterest, Springpad**, and one other of your choice (The Fancy, Fab, Hunch) just to learn what it’s all about. Don’t commit to any one platform. Pinterest may be hot right now, but it’s too early to own this category and some consider the platform of the month a bit one dimensional.
- Take the time to learn what constitutes appropriate and effective conversation strategy on these new platforms. (Hint: it’s not simply about publishing content or adding a Spring This or Pin It button to your site.)
- Look for opportunities to market to the data. We’re a few months or more away from this, but it’s coming.
- Use the platforms yourself. There is no better way to learn and understand their potential.
- If you’re at SxSW this year, come to our panel on the interest graph and deferred intent.
*The Social Graph
A social graph is a digital map that says, “This is who I know.” It may reflect people who the user knows in various ways: as family members, work colleagues, peers met at a conference, high school classmates, fellow cycling club members, friend of a friend, etc. Social graphs are mostly created on social networking sites like Facebook and LinkedIn, where users send reciprocal invites to those they know, in order to map out and maintain their social ties.
*The Interest Graph
An interest graph is a digital map that says, “This is what I like.” As Twitter’s CEO has remarked, if you see that I follow the San Francisco Giants on Twitter, that doesn’t tell you if I know the team’s players, but it does tell you a lot about my interest in baseball. Interest graphs are generated by the feeds customers follow (e.g. on Twitter), products they buy (e.g. on Amazon), ratings they create (e.g. on Netflix), searches they run (e.g. on Google), or questions they answer about their tastes (e.g. on services like Hunch).
Your thoughts? Please share ideas, examples or insights as to where you think things are going.
**Note: In addition to my role as Mullen’s chief innovation officer, I also work as Springpad’s chief marketing officer.
The social filter is hiding a lot of good content.
Discovery engines are converging to first filtering out all content that hasn't been shared, then organizing the remainder by topic. Examples include Pinterest, Svpply, and Paper.li. If something hasn't been pinned, tweeted, liked or otherwise shared, you will not find it through one of these services.
To see how this works, let's trace the path of a single item to a single user.
Say there is a perfect pair of jeans for me. What stands between those jeans and me? They need to be shared by my social network, and they need to match my interests as stated on my profile. In the example of Pinterest, this means that they need to be pinned or liked by someone I follow, and/or onto a board I follow. If I am going to discover them through Pinterest, they cannot get to me until someone pins them. Pinterest is popular, but it's not ubiquitous. There are many good jeans that aren't on Pinterest. So I may not ever discover them. But if someone does share them, they have passed the first hurdle.
The second filter the jeans have to make it through is the interest graph. I have to have expressed interest in this kind of thing. Maybe I follow a men's jeans board on Pinterest. Maybe I follow a men's apparel board.
Then, and only then, can those jeans reach me.
The good thing about this filtering process is that it cuts out low quality and irrelevant content. The bad thing is, it misses a lot of good, relevant content.
There needs to be a way for content that has not been shared to bypass the social filter. Obviously, the content still needs to be vetted somehow. But how? One solution is a very different kind of human filter--editors who judge only whether the content could be possible useful to anyone, that is original, that it is not spam. They don't decide who it is relevant to, just whether it could possibly be useful to anyone. Trapit has an algorithmic discovery engine continually sorting the web into topics. Then, it has human editors for each broad topic (sports, fashion, entertainment) to verify that the content meets a minimum quality standard--that it's not a repost with no original content added (the Huffington Post is notorious for this) or just plain spam.
Now there is a corpus of quality web content that isn't subject to the whims of the crowd. And now the social graph has been cut out of the discovery process. Of course, the social graph will continue to be very important in discovery, but the interest graph can work without it.
I would say the Interest Graph is already here. Also, a bit surprised Google+ did not get a mention when it's the best example of the interest graph in my opinion.
Here's a thought. This is the age of mass customization, especially in communications technologies. The fundamental driver of success as you've noted here will be the contextualization of the brand promise. In the bad old days of direct marketing the focus was audience (targeting), offer ($ incentive) and brand/product. Very often, contextualization took the form of specific ad units developed for specific pubs, ie, cycling, Time, etc. to better position the brand within content-- interest of the pub and its readers. The computing power needed today to give us the clarity we need is nearly here and, once achieved, the overlay of social and interest should reveal a cross sectional opportunity that may allow the same old tenets to apply and become the basis for much more sophisticated execution. I can imagine a day when this contextualization will be optimized with ad units that organically create themselves from a set of brand assets that reside on platform built to anticipate the interests of the likely target audience. So maybe the big difference will be how technology can rapidly optimize delivery of contextualized ad units based on social and interest graphing and then use that contextualization as a method to drive further engagement and power more connections, and ultimately sales, based on these interactions. I think we already have the ability to execute the creative in this way. Someone smarter than me needs to figure out how to use the social and interest data to drive the engine. What do you think?
GuyMastrion timleake I think this makes perfect sense and is closer than we realize.
Edward, thanks for this article. I've been looking for a way to connect content strategy with social strategy and this is the first light at the end of that tunnel for me. The question I have is what tools are you using to gather and analyze Interest graph data?
DezzyDood That is a good question. I'm not really doing much at the moment, but beginning to look at things. By platform it will be easier. For example, if you are a brand using Pinterest or Springpad you have traffic, pins, repins, etc. If you use something like Aditive, the social online ad platform, you have all kinds of robust data from people who share your content with their friends who they know will appreciate. Will look into it and see if I can come up with more. Or email firstname.lastname@example.org @jeffjaner and ask him. He may know.
I'd say that the interest graph is already here. It's called the link graph in SEO. People link to things they like, assign ratings (in the form of links), follow people (in the form of links).
Would you say my definition is too broad and that it needs to be more specific to describe what's happening here?
jeypandian You could argue that, I suppose. We as consumers are obviously looking for what we are interested in. But....do we have an ideal tool to capture, store, save and act on them? Do we get really relevant alerts and information (prices, show times, etc.) when we do? Is it easy to find others with the same interests rather than vice versa? I know what my FB friends like. But what if I just love cycling. How can I find others with same interest who may not be my friends? Even Google Circles needs work. I could put you in my "Interest Graph" circle but you're posting about Italian wines when I check your stream. We're getting there, but it offers way more potential for connecting people and for productive brand engagement.
edwardboches I get what you are saying. Essentially, how do we 'discover' unknown individuals who share the mutual 'interest' and access an information feed from them and known individuals about said 'interest' with no noise.
Am I right?
jeypandian That is one aspect for sure. Other is if you are a brand. Just because someone likes you on FB doesn't mean a lot. But if I have saved or sprung a Canon camera with an intent or desire to buy, or have saved hotels in SF because I plan to travel there, or have captured recipes that I plan to cook, those are all intents with expected actions that brands can use to communicate/incent/motivate me. And I might very well welcome rather than resist their messages and offers. So if they know what I am interested in -- expressed interest and intent -- we both benefit.
Hi Edward, thanks for putting this up... I agree with you on every point. The interest graph calculations add a large layer of computational complexity in the mix but data scientists everywhere are working on this as we speak.
In fact, my company designed a SaaS plaftorm that can calculate and graph the interest graph of any social data feed.Take a look at what we are doing with the crowdfunding hashtag:http://crowdfunding.nexalive.com/
or for that matter the #sxsw hashtaghttp://sxsw-en.nexalive.com/
I would love to have your feedback on it.
cgtheoret Are you searching topic by platform? To see where there is interest? That is interesting in that it identifies topics that matter at the moment (if I understand it.) You should see what Livefyre is up to, as well. They are all over it, also. Good stuff all around.
edwardboches In this case we are taking everything with the SXSW hashtag. We identify concepts with light NLP, and then we perform some metrics to see how and display how all of the concepts interests are connected together. We then put in another layer of calculations to cluster the interests that are most connected together on the actual interest graph, using some basic graph theory metrics.we can do this of course using keyword searches as the inputs and we can use any social data stream... currently the platform interfaces with the Twitter API, Blog Search API's and the FB API.